I ran an experiment on Pavlovia, and it currently shows 36 Submitted.
Yesterday, while data collection was still ongoing, I was able to download a results file via Download results that contained data for approximately 27 participants. However, when I now use Download results from the same experiment page, the downloaded archive is empty. The data folder is present, but it contains no files.
I also checked the GitLab repository. This project currently does not have a data folder, and I could not find any evidence in the commit history that a data folder was ever added. By contrast, other Pavlovia projects that I ran around the same time do have a data folder. I am not sure whether this is related.
Could you please check whether the submitted data for this experiment is still stored on the Pavlovia side, and whether there may be an issue with the generation of the Download results archive or with how the data is being referenced?
Yes, I did switch the experiment to inactive. However, I have been able to download results from my other projects without any issue even after switching them to inactive, so I had assumed that this was not the direct cause. Of course, if this is the intended behavior/specification, I understand.
The experiment data should have been collected between around May 9, 11:30 PM and May 10, 2:00 AM (UTC+9). The final commit was also made at May 9, 7:29 PM, which was definitely before the experiment began. Therefore, I do not think there is a possibility that I committed changes after the experiment and accidentally removed the data folder.
Are you using credits? It would be good to see what the credits tab shows for this experiment, and also the commit history for the period. If you add me (Wake) to the project then I can have a look around to see if I can spot anything unusual.
Yes, I am using credits. I checked the Credits tab, and they appear to have been consumed normally.
It is difficult for me to investigate this on my own, so I would like to invite you to the project. I believe I can add you using your GitLab account. Is the following account the correct one for you? Also, would Reporter permission be sufficient?
Thank you. I have confirmed what you have said and taken some screen shots for my report (e.g. of the commit history). Luckily you have the data for most of your participants.
I’ve added 36 credits to your experiment. At this point please could you test it through yourself to see if your data gets saved to GitLab (i.e. the issue was transient) or not (i.e. the issue is ongoing). In the latter case I would expect to see an error message on screen, but I am guessing none of your participants reported one.
I ran the experiment myself and confirmed that data was added to the data folder in GitLab.
So, just to confirm my understanding: the experimental data was not sent to GitLab for some reason and existed only on Pavlovia, and because I switched the experiment to Inactive while it was in that state, the data was lost. Is that correct?
It is unfortunate that the data for 9 participants cannot be recovered, but I understand that this was partly my mistake as well. I will be more careful about this in the future.
That does seem to be the main issue, I’m afraid. I will update at our end so we can investigate why there might have been a drop in the link between the two servers on that day.
Might your participants be using a VPN to access the experiment?
I notice that the data file starts “undefined” because you have removed “participant” from the expInfo screen. If you don’t want to show your participants the word participant you could edit the code for the data file name
Regarding VPN usage, we did not prohibit participants from using a VPN, so I think it depends on each individual participant. We did not provide or require a VPN on the experiment recruitment side.
I also understand the issue regarding filenames beginning with “undefined”. I will keep this in mind for future reference.
